Lists Home |
Date Index |
At 20:23 4.2.2005, you wrote:
>On Fri, Feb 04, 2005 at 07:10:33PM +0200, Toni Uusitalo wrote:
> > Yes. I tested with the current version - It works. But I think
> > the point still remains - it's not easy to get the DTD parameter
> > entity expansion right ;-)
>One of the changes we made in moving from SGML to XML was in the way
>that parameter entity substitution worked. We said that whitespace
>was added at parameter entity boundaries. This and a number of other
>constraints did reduce the sneaky macro-substitution tricks people
>could play, but macro substitution is still at a syntactic, not a
>semantic level, and that's always a problem when you start wanting
>introspection, reasoning about documents, and processing XML files
>with XML tools...
Yes that whitepace added to the boundaries helps/helped the tokenization
a lot - But still the parameter entity expansion is quite vague part of the
the thing from the parser writers' perspective...
well of course I've this language barrier thing when reading the spec too ;-)
From the DTD writer's perspective macros are very useful.
>If we do go ahead and do future "XML 2.0" work -- and that's by
>no means certain -- I can envision a "DTD module" being outside
>the main core XML spec, and I can also envision DTD deletion, i.e.
>a spec with no explicit document type definition at all.
I guess that's inevitable progress to go and dump the DTDs.
There must be some research going on that measures when people are
ready to switch to the alternatives, I've no clue about this thing myself
usage numbers of DTDs or RelaxNG etc.).